8/30/2019

1

Big Data Neuroscience, Computational Methods, &

Predictive Analytics

Ivo D. Dinov

Statistics Online Computational ResourceHealth Behavior and Biological Sciences

Computational Medicine and BioinformaticsMichigan Institute for Data ScienceNeuroscience Graduate Program

University of Michigan

www.SOCR.umich.edu

Characteristics of Big Neurosci Data

Dinov, et al. (2016) PMID:26918190Dinov (2018) Springer

Example: analyzing observational data of 1,000’s Parkinson’s disease patients based on 10,000’s signature biomarkers derived from multi-source imaging, genetics, clinical, physiologic, phenomics and demographic data elements.

Software developments, student training, service platforms and methodological advances associated with the Big Data Discovery Science all present existing opportunities for learners, educators, researchers, practitioners and policy makers

IBM Big Data 4V’s: Volume, Variety, Velocity & Veracity

BD Dimensions

Tools

SizeHarvesting and management of vast amounts of data

ComplexityWranglers for dealing with heterogeneous data

IncongruencyTools for data harmonization and aggregation

Multi-sourceTransfer and joint-modeling of disparate features

Time Longitudinal data modeling

Multi-scaleMacro to meso to micro scale observations

Incomplete Reliable management of missing data

8/30/2019

2

BD

Big Data Information Knowledge ActionRaw Observations Processed Data Maps, Models Actionable Decisions

Data Aggregation Data Fusion Causal Inference Treatment Regimens

Data Scrubbing Summary Stats Networks, Analytics Forecasts, Predictions

Semantic-Mapping Derived Biomarkers Linkages, Associations Healthcare Outcomes

I K A

Dinov, et al. (2016) PMID:26918190

Predictive Big Data Analytics:Applications to Parkinson’s Disease

Varplot illustrating:

o the critical predictive data elements (Y-axis)

o and their impact scores (X-axis)

AdaBoost classifier for Controls vs. Patients prediction

MLclassifier

accuracy sensitivity specificitypositive

predictive valuenegative

predictive valuelog odds ratio

(LOR)

AdaBoost 0.996324 0.994141 0.998264 0.9980392 0.9948097 11.4882058

SVM 0.985294 0.994140 0.977431 0.9750958 0.9946996 8.902166

Dinov, et al., (2016) PMID:27494614

8/30/2019

3

Acknowledgments

FundingNIH: P20 NR015331, U54 EB020406, P50 NS091856, P30 DK089503, UL1TR002240, R01CA233487

NSF: 1916425, 1734853, 1636840, 1416953, 0716055, 1023115

The Elsie Andresen Fiske Research Fund

Collaborators SOCR: Milen Velev, Yongkai Qiu, Zhe Yin, Yufei Yang, Alexandr Kalinin, Selvam Palanimalai, Syed Husain, Matt Leventhal,

Ashwini Khare, Rami Elkest, Abhishek Chowdhury, Patrick Tan, Pratyush Pati, Brian Zhang, Juana Sanchez, Dennis Pearl, Kyle

Siegrist, Rob Gould, Nicolas Christou, Hanbo Sun, Tuo Wang, Yi Wang, Lu Wei, Lu Wang, Simeone Marino

LONI/INI: Arthur Toga, Roger Woods, Jack Van Horn, Zhuowen Tu, Yonggang Shi, David Shattuck, Elizabeth Sowell, Katherine

Narr, Anand Joshi, Shantanu Joshi, Paul Thompson, Luminita Vese, Stan Osher, Stefano Soatto, Seok Moon, Junning Li, Young

Sung, Carl Kesselman, Fabio Macciardi, Federica Torri

UMich MIDAS/MNORC/AD/PD Centers: Cathie Spino, Chuck Burant, Ben Hampstead, Kayvan Najarian, Stephen Goutman,

Stephen Strobbe, Hiroko Dodge, Hank Paulson, Bill Dauer, Roger Albin, Chris Monk, Issam El Naqa, HV Jagadish, Brian Athey

http://SOCR.umich.edu

Slides Online:

“SOCR News”

http://socr.umich.edu/HTML5/SOCR_TensorBoard_UKBB